On Wed Feb 1 16:56:07 2023 +0000, Rémi Bernon wrote:
I believe this is because of a use-after-free bug that is better handled with the LFH as we avoid touching the blocks now. Currently the `dxgi:dxgi` test crashes silently and only testbot catches that with: ``` The main process has no test summary line The test has unaccounted for failure messages The test has unaccounted for todo messages The test returned success despite having failures ``` That failure is actually due to calling IDXGIResource::GetUsage() with an uninitialized stack parameter. Our stub doesn't initialize the variable. I've sent a patch for this: <https://gitlab.winehq.org/wine/wine/-/merge_requests/2107/diffs>
Not sure why the LFH triggers it, since it should just be uninitialized stack data?
Currently the `dxgi:dxgi` test crashes silently and only testbot catches that with:
``` The main process has no test summary line The test has unaccounted for failure messages The test has unaccounted for todo messages The test returned success despite having failures ```
I don't see this on the testbot patterns page, nor in this specific failure. Where are you seeing this? -- https://gitlab.winehq.org/wine/wine/-/merge_requests/1628#note_22984